Nutritional and endocrine-metabolic aberrations in amenorrheic athletes

Growing evidence suggests that menstrual disturbances in female athletes are related to the metabolic cost of high levels of energy expenditure without compensatory increases in dietary intake. However, the linkage(s) between nutritional deficits and reproductive impairments as a result of slowing of LH pulsatility has not been defined. This study was directed to simultaneously characterize nutritional intake, insulin sensitivity (by rapid iv glucose tolerance test), and 24-h dynamics of insulin/glucose, cortisol, somatotropic [GH/GH-binding protein (GHBP)/insulin-like growth factor I (IGF-I)/IGF-binding proteins (IGFBPs)], and LH axes in highly trained athletes with (cycling athletes; CA) and without (amenorrheic athletes; AA) menstrual cyclicity and in age- and body mass index-matched cycling sedentary controls (CS; n = 8/group). Although daily caloric intake did not differ among the three groups, athletes (CA and AA) consumed less fat and protein than CS. However, the restriction of fat was 50% greater (P < 0.01) in AA than CA and was accompanied by increased carbohydrate (P < 0.05) and fiber (P < 0.01) intake. Athletes, independent of menstrual status, had increased (P < 0.05) insulin sensitivity and reduced insulin levels during the feeding phase of the day. Hypoinsulinemia was more pronounced in AA (P < 0.05) than CA, extending throughout the day, and was accompanied by reduced glucose increments in response to meals (P < 0.05), not seen in CA. Levels of the insulin-dependent IGFBP-1 were markedly elevated (P < 0.001) throughout the diurnal pattern in AA, whereas in CA, a modest elevation (P < 0.001) of IGFBP-1 levels occurred only during the feeding portion of the day. IGFBP-1 levels for the three groups related inversely to 24-h insulin (r = -0.63) and directly to 24-h cortisol (r = 0.69) levels. A 70-80% augmentation (P < 0.001) of 24-h mean GH levels was seen in both groups of athletes, but with distinct pulsatile features. Although pulse amplitude was increased 60% in CA with no change in pulse number, AA displayed more frequent (P < 0.001) pulses, with an elevated (P < 0.01) baseline between pulses. The distorted pattern of GH pulses seen in AA was associated with a 35% decrease in GHBP levels, not seen in CA. Although levels of IGF-I and IGFBP-3 did not differ in either CA or AA, the 2- to 4-fold higher levels of IGFBP-1 in AA than in CA and CS resulted in a 3-fold reduced ratio of IGF-I/IGFBP-1 in AA, which may decrease the bioactivity and hypoglycemic effect of IGF-I. LH pulse frequency was progressively attenuated in the athletes, with a greater (P < 0.001) slowing in AA than CA, unaccompanied by alterations in pulse amplitude or 24-h levels. LH pulse frequency was related positively with insulin (r = 0.65) levels and the ratio of IGF-I/IGFBP-1 (r = 0.69), and negatively with cortisol (r = -0.70) and IGFBP-1 (r = -0.75) concentrations. Stepwise regression analysis suggested that negative influences associated with hypercortisolemia and elevated IGFBP-1 levels predominate in determining GnRH/LH pulsatile activity in these athletes. In sum, although neuroendocrine-metabolic adaptations to the energy cost of exercise training were evident in both groups of athletes, AA displayed alterations distinct from their cycling counterparts, with evidence of a hypometabolic state, including decreased basal body temperature and reduced levels of plasma glucose and serum GHBP, a decrease in the ratio of IGF-I/IGFBP-1, accelerated GH pulse frequency, and elevated interpulse GH levels. Thus, in AA, increased insulin sensitivity, decreased circulating insulin, and a reduced hypoglycemic effect of IGF-I together with elevated GH and cortisol concentrations may comprise a cascade of glucoregulatory adaptations to repartition metabolic fuels for conservation of protein. (ABSTRACT TRUNCATED)